Path: utzoo!mnetor!uunet!lll-winken!lll-lcc!ptsfa!well!perry
From: perry@well.UUCP (Perry S. Kivolowitz)
Newsgroups: comp.sys.amiga
Subject: Re: Recoverable Ram Disks
Message-ID: <4860@well.UUCP>
Date: 27 Dec 87 06:46:32 GMT
References: <2189@crash.cts.com> <887@rocky.STANFORD.EDU> <3005@cbmvax.UUCP>
Reply-To: perry@well.UUCP (Perry S. Kivolowitz)
Organization: Whole Earth 'Lectronic Link, Sausalito, CA
Lines: 27

Incredible. It doesn't matter how many times you explain
things or how well you document things. People still don't
listen. And of all people, Tom Rokicki?

OK, one more time.

Let's say you write to the RRD. The RRD attempts to make a dynamic
(say Dan, can you say Dynamic?) allocation of additional ram but
finds there is no more memory available. It notifies DOS that there
is no more memory left. DOS ignores the result and continues writing
oblivious to the protestations of the RRD.


Later. DOS says...about sector so and so I'd like that back. The ram
disk gives it a freshly formatted sector to which DOS says ``read/write''
error because the checksums don't match up. Delete the file. End of
problem.

You experience this problem because you are violating one or more of the
ground rules I spell out in detail in the doc with the ram disk.

In short one byte still fits only one byte. Declaring an 8 megabyte
rrd in 512k of usuable ram will surely produce erroneous results.
I am sorry, but we haven't succeeded in developing our transporter
beam allowing new ram to materialize in your machine upon
demand.

This is growing tiresome.